EMC: 03cmorley 07master * r1316ae409f98 10/src/hal/components/mux16.comp: Add a mux16 component
EMC: 03cmorley 07master * rb09c95c98b67 10/src/emc/usr_intf/halui.cc: Add direct value input option to overrides
* micges_lucid making rtai
micges_work: thanks for fixing up some of that wiki stuff
I'm going to send out a corrections e-mail about the 2.4.0 announcement. so far I have the typos in the release notes that micges found, and a mention that you probably want to disable the buildbot package repositories
anybody know of something else I should mention?
is it true that many of the sample configs are broken? (NMLFILE)
grep tells me I missed one
configs/classicladder/cl-estop/cl-estop.ini:NML_FILE = emc.nml
ok, I guess they were talking about their own old configs
I don't see the firmware packages under http://linuxcnc.org/hardy/dists/hardy/emc2.4/
how do you get the new hostmot2-firmware-*?
I'm not sure. It's true that you have to EITHER remove NML_FILE from your ini (and you might as well remove emc.nml) OR copy a new emc.nml, because emc.nml changed from 2.3 to 2.4
they're in /base/ because they work with 2.3 or with 2.4.
oh ok, I didn't look back far enough.
jepler: wait few hours, I'll look at this at home
maybe I'll found some more typos
I sure need to update my machines
I put my notes I have so far for lucid on the wiki.
I'm still having boot problems on some computers with the latest rtai patches. I'm going to try disabling SMP in the config next and see if that helps. I did that with 9.10 and it didn't seem to make a difference - so we'll see.
If anyone sees any kernel settings that need to be enabled or disabled that I don't have in my notes, please add them.
oh, looks like someone found my notes and linked to them on the front page!
seb_kuzminsky: are you flying or driving to cnc workshop?
seb_kuzminsky: I'm asking because jeff and I will have a few extra days and plan to do some other stuff near the area
i gotta get back to my family... squeezing in a handful of days for the workshop is tough enough...
if I don't modify VERSION or debian/changelog, update-dch-from-git makes a package name like 2.4.0-1-gabcdefg which will be greater than 2.4.0 but less than 2.4.1. Is that good enough to let the buildbot 2.4 packages be reenabled?
mozmck_work: is there anything obvious in common between the machines you've tried that won't boot?
right now I'm just trying on two machines. Both are intel P4, one that boots about half the time uses onboard intel video, the other has an nvidia card.
yuck, half the time?
what's it do when it doesn't boot?
The one with nvidia card doesn't want to boot at all. If I disable modesetting it will boot into a terminal, but can't start X because it can't find screens or sometihng.
black screen with blinking cursor immediately after grub.
have you tried to rebuild the stock kernel without any new patch at all?
hit the power button, power back up and it will generally boot fine then.
no, I figured that would be the same as the one from ubuntu. and it takes 1.5 or 2 hours to do a build on that machine.
their setup that loads the (X?) video drivers early for that splash screen thing has got to be complicated. I wonder if it's not a problem with rtai at all.
I think the splash screen is using KMS and not X drivers. I think it is a problem with KMS and rtai conflicting.
you can get the grub menu by holding a shift key during boot
I'm sure you know much more about it than I do - I was just thinking I'd be tempted to try to rebuild a totally unmodified setup and see if even that works.
I found this out when installing nvidia drivers prevented my machine from booting correctly :)
(luckily I had a smartphone with me)
mozmck_work, does the system with the nvidia card also have onboard video?
well the stock ubuntu kernel loads fine, and I'm using the ubuntu git repository for my builds.
SWPadnos: I have a multiboot setup and the grub menu comes up automatically, but that's good to know for single boot setups.
The nvidia machine does have onboard video. Crazy thing is this machine boots fine from a slightly older ubuntu kernel and magma patch, and gets good latency readings.
I was a bit scared when ESC didn't work :)
I don't know if it's still there, but there was an issue with nvidia drivers in the beta stage for 10.04
mozmck_work: are you building very latest kernel version with rtai?
latest ubuntu kernel. 2.6.32-22.34
yeah, this machine with the nvidia card boots fine with 2.6.32-20-rtai (my build of course), and I think the magma patch was for 220.127.116.11
jepler: i agree that's the version number it'll get, and that that version number will work well
same kernel config... paolo said the latest patch (for 18.104.22.168) had some boot fix-ups, but maybe I should try the latest kernel with the older patch.
i have not disabled the buildbot in any way, the only thing i did was move the "1:2.4.0" packages out of the way so folks apts won't see them
next commit on 2.4 will produce a package like you said, and ppl will be offered the update
i *think* that's what we want...
yes; if people want to stay on the bleeding edge of 2.4 development, we'll benefit from that
.. as long as that's what they really wanted, which is why I'm going to mention it in a follow-up to the announcement
mozmck_work: can you check if your wiki page about rtai comiling is up to date?
I have hard time to follow it, but I've managed to compile stock ubuntu
EMC: 03seb 07master * r72717505d2b3 10/debian/update-dch-from-git: don't prepend to deb changelog if it's up-to-date
micges: I don't think I put anything under rtai yet. I think my stuff on the kernel is up to date though. I use it each time I do something because I forget stuff...
ok, I was doing it first time, I'll check it again tomorrow
some of this I got in different places, but I had to figure some out myself. They changed file locations for some things between karmic and lucid, and a lot of the information you find elsewhere is wrong (says the ubuntu kernel folks!).
I've noticed ;)
(of course, some of mine might be as well)
If you can make some of it easier to follow that would be great. I'll be refining it along the way too.
I'll try tomorrow
jepler: if we merge that commit into 2.4, then when you release 2.4.1 the buildbot should make the correct deb for you, & the release manager can just copy it to the deb archive on linuxcnc.org
seb_kuzminsky: that's an interesting possibility
jepler: I didn't found any more typos on wiki
micges: thanks for reading it over carefully
hey guys, should rtai/emc2 beable to work in parallels? On my laptop I compiled rtai and emc without problems, but within parallels I keep getting : Pid: 2675, comm: insmod Not tainted (22.214.171.124-rtai-rvt #1) Parallels Virtual Platform EIP: 0060:[<d08fad8a>] EFLAGS: 00010246 CPU: 0 ..........
what is parallels?
cradek: it's a virtualization software; I think it may be for mac
you can't control a machine that way, and if you just want to run the simulator, you can build in sim mode without rtai.
there is no point whatsoever to having rtai inside a virtual machine
I see you built your own kernel; do you know that the kernel is good? Does it work on a non-virtualized PC?
did anyone considered xenomai as rtai replacement?
aystarik: I don't know of anybody who's looked into it
I am running it's latency test on 64 bit machine, it gives me 2.65 us without reserving CPU...
its own latency test? or do you have emc running on it?
no, I wanted to see if it's worth the trouble...
darn, my advice didn't get Neil running :(
aystarik: someone on the mailing list named Neil is having trouble with 2.4.0 on his system.
what's the problem of stable release... No one cares about development branch until it becomes 'stable'...
aystarik: even worse, when people won't use any .0 release
aystarik: (I'm not sure any active developers even have any ppmc hardware - that makes it pretty hard to test)
cradek: with linux kernel I did not happen to hit .0 release :) started around .12 :)
jon elson seems to be active at cnczone, no?
no idea, I don't read cnczone
I have a USC, but no EMC2 machine at the moment
actually, I think I could test shortly, I have a machine with EMC2/RTAI and I think it even has a parallel prot
on my system without a ppmc, I get this: [ 333.496528] PPMC: ERROR: no boards found on bus 0, port 0378
which is past the error Neil is getting
ppmc shouldn't use PARPORT at all, so there's clearly something else going on
ppms uses a parallel port, but it doesn't use the (linux) parport driver
in 2.4 it does
oh. then nevermind me
geez. I really need to catch up
in 2.4 all the parport drivers (dumb parport, pluto, 7i43, ppmc/usc) can use the linux-detected devices cooperatively, or they can do it the old way
cradek: we're maintain one gui and barely all rest
and still there are people out there creating new guis
in pascal... :(
I'm writing my next gui in lispqt
:) why not cobol?
my will be in d
D does look interesting. I keep hearing more of it - is it growing?
apl + google web toolkit + Flash
[19:50:42] <andypugh> http://en.wikipedia.org/wiki/Brainfuck
mozmck_work: yes it's going to be fairy adult soon
fairy? or fairly
ah, thought so. I've downloaded D and played with it a little, and it looks promising.
A toolkit I've used a little now that seems promising is fltk.
I still like python + gtk
I'm not very used to python yet. Still do mostly C/C++
I don't see why anyone would develop for a language without an open-source compiler or interpreter and runtime... I see that D now has an open source compiler (gdc, based on gcc) so perhaps I can be less than totally dismissive of it..
(or is it? the gdc page refers to needing the dmd front-end which as far as I could see on digitalmars.com is not open source)
gdc works, but they say it's not as good as dmd yet
I thought it was fully open source, but I may be wrong. That probably would help D a lot though to have a really good open source compiler.
One thing I can't understand is the push for C# on linux. Seems like anything using Mono is really slow too.
whether it's Java, C#, D, or some other language, I can't blame developers for continuing to want something better than C++.
> Note: all D users agree that by downloading and using D, or reading the D specs, they will explicitly identify any claims to intellectual property rights with a copyright or patent notice in any posted or emailed feedback sent to Digital Mars.
that's hiding down at the bottom of http://www.digitalmars.com/d/
can you say "desert island test"?
What does that paragraph actually mean? I thought I understood english.
they claim that if you don't do an explicit thing, you implicitly are giving up rights that you normally automatically have
Which sounds to include reading the documentation?
only relevant if you post or send feedback to Digital Mars.
or care about such things
I think a lot of forums claim copyright of everything you post.
well that's asinine too, if so
I was interpreting it in yet a different way
namely, if you use the D language or read the D specs, AND you are aware of copyright or patent infringement, you MUST contact Digital Mars about it
ah, that could be too.
I'm with andypugh now about not being able to parse it
except for the word "any" in front of "posted or emailed..." Not clearly worded for sure.
often, people use poorly-thought-out language to express poorly-thought-out ideas
(wait, is my prejudice showing?)
I appreciate that they want to have the opportunity to fix any IP issues in their software, and they can't do that without being aware of them.
the 'any' causes a parse error when I try to read it the way jeff said
maybe I'm all wrong, then
yeah, that's what I was trying to say.
heh, not that you were all wrong jepler! but what cradek said.
and it doesn't specify that they mean IP claims related to D
cradek: well, no
so are they saying that if I e-mail them, I'd better have one of those lengthy disclaimers at the bottom
I still think it says if you post [where?] or email them, and you don't have an explicit statement of copyright, you implicitly renounce it
... anytime after you've read [some of?] their documentation
or at least allow them to use it without asking you further.
I guess I don't care about the dissection of this bit of legalese that much
except I can't help but observe that it's not asking for identification of claims relating only to the message
[20:49:25] <jepler> http://www.wired.com/magazine/wp-content/images/ff_radioshack/ff_radioshack10_f.jpg
-- only 3x as much as an iphone!
I miss full-bandwidth analog phones that were big enough to put the microphone in front of your mouth where it actually works
I miss my bag phone that had great reception anywhere
for another $69.95 you could get a data interface for tandy laptop computers (the model 100 I imagine)
15 hours standby / 1.5 hours talk
that's where our phones are actually better today
my bag phone stayed plugged into my truck battery all the time
analog pocket size phones had a hard time lasting through an 8 hr work day on a battery
you needed a second battery for the evening, and then it was hard to charge both again by morning
I have a friend with an Android G1 which is about that bad.
did the "edit tool table" thingy get tossed? I see it defaults to vim now
JT-Dev: depends on the config
I'm running the axis 2.4/5 sim
I guess someone changed it then...
* JT-Dev goes to look at the ini for it
Is comp the "preferred" way to code standalone HAL modules?
hmm I don't see an entry in the ini for tooleditor
greppping I find axis.py:tooleditor = inifile.find("DISPLAY", "TOOL_EDITOR") or "tooledit"